Sebastian Thrun and Chris Anderson: What AI is -- and isn't
Sebastian Thrun et Chris Anderson: La nouvelle génération d'ordinateurs se programme seule
Sebastian Thrun is a passionate technologist who is constantly looking for new opportunities to make the world better for all of us. Full bioChris Anderson - TED Curator
After a long career in journalism and publishing, Chris Anderson became the curator of the TED Conference in 2002 and has developed it as a platform for identifying and disseminating ideas worth spreading. Full bio
Double-click the English transcript below to play the video.
ce qu'est le « machine learning »
what machine learning is,
and also of the concern
de toutes ces craintes
intelligence and machine learning
artificielle et le « machine learning »
in its past until recently.
de gloire jusqu'à récemment.
of computing and datasets
de calcul et de volume de données
rendre les machines intelligentes.
say, your phone,
disons, votre téléphone,
des ingénieurs logiciels
very long kitchen recipe,
longue recette de cuisine
turn down the temperature.
baisse la température.
augmente la température. »
the temperature."
has 12 million lines of code.
12 millions de lignes de code.
de lignes de code.
can cause your computer to crash.
provoquer une panne de votre ordinateur.
makes so much money.
gagne autant d'argent.
can find their own rules.
peuvent trouver leurs propres règles.
deciphering, step by step,
qui écrit, pas à pas,
the computer examples
à l'ordinateur des exemples
which recently was won by Google.
qui a récemment été acheté par Google.
you would really write down all the rules,
vous voulez écrire toutes les règles,
d'un million de parties
déduire ses propres règles
residing Go champion.
mondial du jeu de go.
à faire appel à des ingénieurs logiciels
the software engineer
vers les données.
where this has become really possible --
où cela est vraiment devenu possible --
était sur le « machine learning ».
was about machine learning.
ne la lisez pas,
insignificant, don't read it,
aussi gros que le cerveau d'un cafard.
were as big as a cockroach brain.
to really emulate
pour imiter
take advantage of the fact
plus de données que les gens.
much more data than people can.
more than a million games.
plus d'un million de parties.
study a million games.
étudier un million de parties.
a hundred billion web pages.
de milliards de pages web.
a hundred billion web pages.
centaine de milliards de pages.
the computer can find rules
peut trouver des règles
to, "If he does that, I will do that,"
comme « S'il fait cela, je fais ceci »,
looks like a winning pattern,
« Voici à quoi ressemble
a winning pattern."
on élève les enfants.
how you raise children.
aux enfants une règle pour chaque cas
giving kids a rule for every contingency
avec ce grand programme.
and they have this big program.
ils sont giflés ou fessés,
they get slapped or spanked,
une bonne note à l'école,
a good grade in school,
so much easier all of a sudden.
beaucoup plus facile tout à coup.
suffit d'avoir beaucoup de données.
We just give them lots of data.
de l'amélioration spectaculaire
to the spectacular improvement
comment cela fonctionne ?
véhicule autonome
un spin-off appelé Voyage.
into a spin-off called Voyage.
l'apprentissage profond
called deep learning
from Mountain View, California,
Mountain View, en Californie,
and 133 traffic lights.
et 133 feux de circulation.
véhicule autonome de Google
the Google self-driving car team.
meilleurs ingénieurs logiciels mondiaux
the world's best software engineers
dans le cerveau de l'ordinateur,
into the computer brain,
that often surpasses human agility.
qui surpasse l'agilité humaine.
about 33 miles, an hour and a half.
une heure et demie.
of this program on the left,
partie de ce programme à gauche,
the computer sees as trucks and cars
comme les camions, les autos
et ainsi de suite.
image, which is the main input here,
caméra, qui est la source principale ici,
other cars, traffic lights.
les feux.
to do distance estimation.
estimer les distances.
in these kind of systems.
dans ce genre de systèmes.
and so on depicted by the laser.
les arbres représentés par le laser.
sur l'image de la caméra.
is centering on the camera image now.
comme les radars et les lasers
sensors like radars and lasers
à très bas coût.
on the left thing, what is that?
qu'est-ce que c'est ?
pour le régulateur de vitesse,
for your adaptive cruise control,
comment réguler la vitesse
how to regulate velocity
the cars in front of you are.
de la voiture devant vous.
got an example, I think,
un exemple, je pense,
learning part takes place.
d'apprentissage a lieu.
proposé aux étudiants d'Udacity
a challenge to Udacity students
de véhicule autonome.
a self-driving car Nanodegree.
how to steer this car?"
comment conduire cette auto ».
to get the steering right.
quasi-impossible de conduire correctement.
"It's a deep learning competition,
d'apprentissage profond,
like Google or Facebook,
telle que Google et Facebook,
at least six months of work.
au moins six mois de travail.
100 submissions from students,
100 soumissions d'étudiants,
parfaitement réussi.
le faire sur ces images,
drive on this imagery,
to a computer now,
de données
to comprehend the data,
pour comprendre,
of powerful applications
d'applications puissantes
the other day about cancer.
CA: This is cool.
CA : Merci.
into what's happening
de ce qui se passe
400,000 dollars a year,
400 000 $ par an,
to be a good dermatologist.
un bon dermatologue.
the machine learning version of it.
« machine learning ».
for these machine learning algorithms.
pour ces algorithmes d'apprentissage.
by a Facebook Fellow called Yann LeCun,
camarade de Facebook appelé Yann LeCun,
as the human brain.
à un cerveau humain.
but it emulates the same thing.
mais il simule la même chose.
the visual input and extracts edges
le visuel et extrait les contours
more complicated edges
plus complexes,
really complicated concepts.
construire des concepts très compliqués.
cat faces and dog faces
des têtes de chat et de chien
at Stanford has shown is that
Stanford ont montré, est que
of skin conditions,
de maladie de la peau,
that this is the case,
that we presented to our network
que l'on a présenté à notre réseau
Stanford-level dermatologists,
du niveau de Stanford
the performance classification accuracy
performances de classifications
That's a moving piece.
C'est émouvant.
in "Nature" earlier this year
« Nature » cette année, la comparaison
dermatologists images
entre des dermatologues
we had the correct classification.
assurer une classification correcte.
nos collaborateurs à Stanford.
by one of our collaborators.
c'est que ce collaborateur,
one of the three best, apparently,
l'un des trois meilleurs, apparemment,
« Ce n'est pas un cancer. »
"This is not skin cancer."
a second moment, where he said,
and ran our piece of software,
et a lancé notre logiciel,
the iPhone a little bit more than myself,"
confiance dans l'iPhone qu'en moi-même »,
to get it biopsied.
pour une biopsie.
un mélanome agressif.
que l'on a effectivement trouvé,
that we actually found,
d'apprentissage profond,
would have gone unclassified,
qui aurait été non diagnostiqué,
l'apprentissage profond.
for an app like this right now,
pour une telle application,
making an app that allows self-checking?
cet auto-diagnostic disponible ?
de questions à ce sujet,
about cancer apps,
de personnes,
10, 15, 20 melanomas removed,
10, 15 ou 20 mélanomes enlevés,
might be overlooked, like this one,
comme celui-ci, passe inaperçu,
ou de participation à des conférences.
these days, I guess.
de plus de tests.
and impress a TED audience.
tape-à-l’œil pour impressionner TED.
something out that's ethical.
quelque chose d'éthique.
l'assistance d'un médecin
the assistance of a doctor
et que les données se tiennent,
and our data holds up,
ce type de technologie
to take this kind of technology
Stanford ne sont jamais allés.
doctors never, ever set foot.
cette armée d'étudiants d'Udacity,
with this army of Udacity students,
de « machine learning »
a different form of machine learning
with a form of crowd wisdom.
automatique et de sagesse de foule.
that could actually outperform
vous pouvez surpasser
même une très grande ?
even a vast company?
qui m'époustouflent,
instances that blow my mind,
is these competitions that we run.
les compétitions que nous organisons.
a self-driving car
San Francisco sur la route.
to San Francisco on surface streets.
après 7 ans de développement,
after seven years of Google work,
and three months to do this.
et trois mois pour faire cela.
an army of students
une armée d'étudiants
le « crowdsourcing ».
who use crowdsourcing.
des bugs en « crowdsourcing »
where people do bug-finding crowdsourcing
pour toutes sortes de projets.
in crowdsourcing.
this car in three months,
cette voiture en trois mois,
who are never hired,
qui ne sont jamais embauchées,
and I don't even know.
je ne le sais même pas.
maybe 9,000 answers.
peut-être 9 000 réponses.
peut-être pas la meilleure chose à faire.
which is maybe not the best thing to do.
of their education, too, which is nice.
une part de leur formation.
de produire des résultats extraordinaires
to produce amazing deep learning results.
et apprentissage machine est incroyable.
and great machine learning is amazing.
lors du premier jour [de TED2017]
the first day [of TED2017]
avérés être deux joueurs amateurs
turned out to be two amateur chess players
médiocres ou médiocres-à-bons
mediocre-to-good, computer programs,
with one great chess player,
you're talking about a much richer version
d'une version bien plus riche
le fantastique panel hier matin,
the fantastic panels yesterday morning,
et la réponse humaine,
that we sometimes confuse
with this kind of overlord threat,
avec cette sorte de menace de domination,
consciousness, right?
is for my AI to have consciousness.
est une IA consciente.
with the dishwasher
amoureux du lave-vaisselle,
je ne suis pas gentil,
and I don't want them.
et je n'en veux pas.
an augmentation of people.
d'augmenter les capacités.
de Kasparov était très juste.
of human smarts and machine smarts
de machines intelligents
is as old as machines are.
forts est aussi vieille que les machines.
place because it made steam engines
avec des machines à vapeur,
ne sachant pas cultiver seuls,
that couldn't farm by itself,
mais rendus plus forts.
it made us stronger.
will make us much, much stronger
d'IA nous rendra beaucoup plus forts
of this for some people,
qui effraie certains,
scary for people is when you have
réécrire son propre code
rewrite its own code,
multiple copies of itself,
copies de lui-même,
if a goal is achieved and improved.
si le but est atteint et amélioré.
on an intelligence test.
sur un test d'intelligence.
that's moderately good at that,
de versions de cela.
d'effet d'emballement
some sort of runaway effect
on Thursday evening,
on Friday morning,
le vendredi matin,
of computers and so forth,
des ordinateurs,
what I heard you say.
vous venez de dire.
we had exactly this thing:
nous avons exactement cela :
the game against itself
is a rewriting of the rules.
une réécriture des règles.
absolutely no concern
absolument aucun risque
these are all very single-domain things.
ce sont des systèmes mono-domaines.
that seemed nearly capable
qui semble presque capable
and understand in the sense that we can,
comprendre comme on le fait,
patterns of meaning.
qui ont du sens.
as this broadens out,
l'évolution de la technologie,
effet d'emballement ?
kind of runaway effect?
la frontière, honnêtement.
I draw the line, honestly.
I don't want to downplay it --
je ne veux pas le minimiser,
et pas d'actualité pour le moment
the thing that's on my mind these days,
la grande révolution est ailleurs.
is something else.
Artificielle à ce jour
to the present date
is because of massive numbers of Go plays,
l'immense nombre de parties de go jouées
or fly a plane.
ni piloter un avion.
or the Udacity self-driving car
ou celle d'Udacity
et elle ne peut rien faire d'autre.
and it can't do anything else.
à un domaine donné,
domain-specific function,
on this thing called "general AI,"
appelé « IA générale »,
"Hey, invent for me special relativity
« Invente une théorie sur la relativité
and I want to acknowledge them.
et je veux les reconnaître.
"What if we can take anything repetitive
je pouvais prendre une tâche répétitive
100 times as efficient?"
we all worked in agriculture
on était tous agriculteurs,
des tâches répétitives.
travaillent au bureau
des singes de tableurs.
doing repetitive things,
des tâches répétitives,
des tâches répétitives.
of being able to take an AI,
l'IA va pouvoir nous aider,
as effective in these repetitive things.
efficaces pour ces tâches répétitives.
a little terrifying to some people,
un peu terrifiant pour certains,
can do this repetitive thing
peut faire des choses répétitives
is the thing that's talked about
puisqu'on en parle beaucoup
d'emplois disparaissent,
glorious aspects of what's possible.
glorieux de ce qui est possible.
and it's a big issue,
un problème important
by several guest speakers.
par plusieurs orateurs.
optimistic person,
un discours optimiste,
back 300 years ago.
300 ans en arrière.
of continuous war,
à 140 ans de guerre ininterrompue,
n'existent même pas,
or software engineer or TV anchor.
ingénieur logiciel ou présentateur TV.
with a little steam engine in his pocket,
machine à vapeur dans la poche,
as strong, so you can do something else."
et vous libérer du temps. »
il n'y avait pas de podium,
there was no real stage,
with the cows in the stable,
dans l'étable avec les vaches,
concerned about it,
et si la machine le fait pour moi ? »
and what if the machine does this for me?"
je le mentionne est...
past progress and the benefit of it,
passé et ses bénéfices,
or electricity or medical supply.
l'électricité ou les médicaments.
which was impossible 300 years ago.
ce qui était impossible il y a 300 ans.
the same rules to the future.
les mêmes règles pour le futur.
travail en tant que PDG,
of my work is repetitive,
mon travail est répétitif,
on stupid, repetitive email.
sur des courriels répétitifs stupides.
that helps me get rid of this.
qui m'aide à me débarrasser de ces tâches.
are insanely creative;
nous sommes follement créatifs,
more than anybody else.
plus que n'importe quelle autre.
I think you can go to your hotel maid
aller voir votre femme de chambre
you find a creative idea.
vous trouverez une idée créative.
is to turn this creativity into action.
transformer cette créativité en action.
build Google in a day?
construire Google en un jour ?
inventer le prochain Snapchat,
and invent the next Snapchat,
in my opinion.
great side effects.
de grands effets secondaires.
and education and shelter
les médicaments, l'éducation, le logement
affordable to all of us,
that this time it's different
cette fois c'est différent
that we've used in the past
qui prennent en charge cela,
is that, not completely,
different from the kind of creativity
différent de la créativité
belief as an AI person --
en tant qu'expert de l'IA --
any real progress on creativity
de progrès en créativité
important pour les gens de le réaliser,
really important for people to realize,
artificielle » est alarmant
intelligence" is so threatening,
nous jette dans un film,
tossing a movie in,
domine l'humanité,
the computer is our overlord,
à faire des tâches répétitives.
do repetitive things.
ces tâches répétitives.
entirely on the repetitive end.
de radios des poumons.
menace pour l'humanité.
we've become superhuman.
the Atlantic in 11 hours.
à la nage en 11 heures.
de notre poche
shouting back to us.
la personne peut nous répondre.
We're breaking the rules of physics.
Nous brisons les règles de la physique.
we're going to remember everything
nous nous rappellerons de tout
in my early stages of Alzheimer's.
avec mes débuts d'Alzheimer.
an IQ of 1,000 or more.
avoir un QI de 1 000 et plus.
spelling classes for our kids,
d'orthographe pour nos enfants,
d'orthographe.
is that we can be super creative.
que l'on va devenir super créatifs.
it's going to be painful,
de plus que ces emplois.
of more than those jobs.
to just a new level of empowerment
s'élever à un nouveau niveau d'autonomie
60-100,000 years old, give or take --
à peu près, 60-100 000 ans,
in terms of invention,
chérir comme inventions,
qu'on a créées,
it's a little bit older.
qui sont un peu plus anciens.
manufacturing, penicillin --
moderne, la pénicilline --
dans les 150 prochaines années
has gone up, not gone down, in my opinion.
augmenté et non diminué, selon moi.
things have been invented yet. Right?
intéressantes ont déjà été inventées.
encore. J'espère que je vais changer ça.
Hopefully, I'll change this.
people laughed about. (Laughs)
(Rires).
secrètement sur les voitures volantes.
Working secretly on flying cars.
plus longtemps.
implant in our brain
magique dans le cerveau qui
quand vous l'aurez !
once you have it, you'll love it.
we haven't invented yet
from one location to another.
that flight wouldn't exist,
qu'on ne pourrait pas voler,
than you could run,
que la vitesse de course,
that you can't beam a person
qu'on ne peut pas téléporter une personne
and your brilliance.
et votre génie.
ABOUT THE SPEAKERS
Sebastian Thrun - Educator, entrepreneurSebastian Thrun is a passionate technologist who is constantly looking for new opportunities to make the world better for all of us.
Why you should listen
Sebastian Thrun is an educator, entrepreneur and troublemaker. After a long life as a professor at Stanford University, Thrun resigned from tenure to join Google. At Google, he founded Google X, home to self-driving cars and many other moonshot technologies. Thrun also founded Udacity, an online university with worldwide reach, and Kitty Hawk, a "flying car" company. He has authored 11 books, 400 papers, holds 3 doctorates and has won numerous awards.
Sebastian Thrun | Speaker | TED.com
Chris Anderson - TED Curator
After a long career in journalism and publishing, Chris Anderson became the curator of the TED Conference in 2002 and has developed it as a platform for identifying and disseminating ideas worth spreading.
Why you should listen
Chris Anderson is the Curator of TED, a nonprofit devoted to sharing valuable ideas, primarily through the medium of 'TED Talks' -- short talks that are offered free online to a global audience.
Chris was born in a remote village in Pakistan in 1957. He spent his early years in India, Pakistan and Afghanistan, where his parents worked as medical missionaries, and he attended an American school in the Himalayas for his early education. After boarding school in Bath, England, he went on to Oxford University, graduating in 1978 with a degree in philosophy, politics and economics.
Chris then trained as a journalist, working in newspapers and radio, including two years producing a world news service in the Seychelles Islands.
Back in the UK in 1984, Chris was captivated by the personal computer revolution and became an editor at one of the UK's early computer magazines. A year later he founded Future Publishing with a $25,000 bank loan. The new company initially focused on specialist computer publications but eventually expanded into other areas such as cycling, music, video games, technology and design, doubling in size every year for seven years. In 1994, Chris moved to the United States where he built Imagine Media, publisher of Business 2.0 magazine and creator of the popular video game users website IGN. Chris eventually merged Imagine and Future, taking the combined entity public in London in 1999, under the Future name. At its peak, it published 150 magazines and websites and employed 2,000 people.
This success allowed Chris to create a private nonprofit organization, the Sapling Foundation, with the hope of finding new ways to tackle tough global issues through media, technology, entrepreneurship and, most of all, ideas. In 2001, the foundation acquired the TED Conference, then an annual meeting of luminaries in the fields of Technology, Entertainment and Design held in Monterey, California, and Chris left Future to work full time on TED.
He expanded the conference's remit to cover all topics, including science, business and key global issues, while adding a Fellows program, which now has some 300 alumni, and the TED Prize, which grants its recipients "one wish to change the world." The TED stage has become a place for thinkers and doers from all fields to share their ideas and their work, capturing imaginations, sparking conversation and encouraging discovery along the way.
In 2006, TED experimented with posting some of its talks on the Internet. Their viral success encouraged Chris to begin positioning the organization as a global media initiative devoted to 'ideas worth spreading,' part of a new era of information dissemination using the power of online video. In June 2015, the organization posted its 2,000th talk online. The talks are free to view, and they have been translated into more than 100 languages with the help of volunteers from around the world. Viewership has grown to approximately one billion views per year.
Continuing a strategy of 'radical openness,' in 2009 Chris introduced the TEDx initiative, allowing free licenses to local organizers who wished to organize their own TED-like events. More than 8,000 such events have been held, generating an archive of 60,000 TEDx talks. And three years later, the TED-Ed program was launched, offering free educational videos and tools to students and teachers.
Chris Anderson | Speaker | TED.com